Five-Minute Field Notes: a local Gemma card that sends me outdoors
This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass What I Built Five-Minute Field Notes makes a tiny observation mission for a place I can already access: a park, garden, quiet sidewalk, campus, or balcony. I select what I want to notice, whether I will stroll or stay seated, and whether I have 5, 10, or 15 minutes. A local model creates three short cues and a…
This is an account of a lone submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass. The participant created a tool called Five-Minute Field Notes. This app generates a compact observation mission tailored to a nearby location, be it a park, garden, quiet sidewalk, campus, or balcony.
Users specify the duration of their observation - five, ten, or fifteen minutes - and whether they intend to stroll or remain seated. The app then comes up with three brief prompts and a reflective question upon the user's return. However, it does not provide maps, real-time weather, or species identification.
The application is a solo project without any credited collaborators. A brief demo illustrates the running app, user selections, the local model producing a card, and the final printable card. No voice or background music is included.
The model used for this task is Gemma 3 1B, a local model that runs on the user's Mac after a one-time internet download. The app's interface runs on a responsive HTML interface and printable card. The core functionality lies in the local Python server conversing with Ollama, Google's open-weight Gemma 3 1B model.
The app's two primary files are app.py, which handles the local Python server and Ollama integration, and index.html, which holds the interactive interface and printable card. The README.md file offers setup instructions, demonstration flow, model request validation, and card verification.
The app is designed to run locally without the need for constant internet connection once the model is downloaded. Users can inspect and modify the prompts, swap in different models using an environment variable, and operate the app without any per-request charges. This makes the project a small experiment that can be adapted and repurposed by others according to their own settings and languages.
However, a note of caution is provided: a 1B model may occasionally produce repetitive or awkward cues, and users are advised to apply their judgment and remain in safe, accessible areas.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
